nowfound

Life & fun · November 12, 2026

RealtimeVAD – trading false positives for 110ms earlier end-of-speech

Faster speech off signals for spec gen voice pipelines - KoljaB/RealtimeVAD

RealtimeVAD – trading false positives for 110ms earlier end-of-speech launched on November 12, 2026 with 33 votes, #412 of 3,085 launches that month and more than 95% of that year's launches.

Visit github.com ↗Alternativestop 13% of November 2026

In plain words

RealtimeVAD is a voice activity detection tool designed for speculative generation voice pipelines. It detects when speech ends approximately 110 milliseconds earlier than standard approaches, prioritizing speed over perfect accuracy. The tradeoff accepts more false positives to enable faster speech-off signals, making it useful for real-time voice applications where lower latency is critical. The project is available on GitHub for developers building voice-based systems.

written from the facts on this page · November 2026

Does a similar job

all alternatives →
  • I built a sub-500ms latency voice agent from scratchMar 2026 · ntik.me · ▲570

    I built a voice agent from scratch that averages ~400ms end-to-end latency (phone stop → first syllable). That’s with full STT → LLM → TTS in the loop, clean barge-ins, and no precomputed responses. What moved the needle: Voice is a turn-taking problem, not a transcription problem. VAD alone fails; you need semantic end-of-turn detection. The system reduces to one loop: speaking vs listening. The two transitions - cancel instantly on barge-in, respond instantly on end-of-turn - define the experience. STT → LLM → TTS must stream. Sequential pipelines are dead on arrival for natural…

  • Real-time AI Voice Chat at ~500ms Latency2025 · github.com · ▲524

    Have a natural, spoken conversation with AI! Contribute to KoljaB/RealtimeVoiceChat development by creating an account on GitHub.

  • Conversational AI 2.0 From ElevenLabs2025 · elevenlabs.io · ▲652

    Luke from ElevenLabs here. We just released Scribe Realtime v2, our new real-time Speech to Text model. We rebuilt it from the ground up around a new architecture that uses predictive transcription – it anticipates the next likely word before it’s spoken. That makes it both fast (around 150 ms latency) and highly accurate. Scribe v2 Realtime outperforms every low latency transcription model across 30 commonly used EU and Asian languages – 93.5% accuracy, compared to Gemini 2.5 Flash (91.4%), GPT-4o MiniTranscribe (90.7%), and Deepgram Nova 3 (85%). It supports 90+ languages including…

  • Outspeed2024 · ▲462

    Platform for realtime voice and video AI

  • gpt-realtime-1.5 by OpenAIFeb 2026 · developers.openai.com · ▲300

    Tighter instruction adherence in speech agents

  • Tempo2024 · gettempo.app · ▲387

    Stop speaking too fast during presentations

More life & fun this month

the category →

Launched alongside, November 2026

the whole month →